Human - Artificial Intelligence Teaming for Automotive Applications: A Review

Human Artificial Intelligence Teaming (HAIT) is a significant topic that is dominating different research domains. One of these domains is the automotive industry, whereby automation is suggested to certain aspects of driving, while the driver can intervene and be aware of the decisions. Trust is a major issue; hence the AI collaborates with the human towards making a decision regarding different aspects of driving. The Internet of Vehicles (IoV) is a topic that can use HAIT in many of its applications. A major point of the HAIT application is the increase in the transparency of the AI process and trust is being built between the two teammates. In this paper, the goal is to offer a comprehensive review of HAIT and its significance, going deep into various representations to facilitate the development of automated vehicles systems. HAIT seeks to promote trust in automated automotive systems, particularly regarding data sourced from vehicle sensors. The human roles 'in,' 'on,' and 'over' the loop within HAIT is provided, elucidating their pivotal contributions. Furthermore, ongoing academic contributions are reviewed integrating HAIT into the automotive sector, emphasizing the symbiosis between IoV and AI to forge unified solutions. The solutions have been separated according to their functionality and models used comprising Reinforcement Learning, Hidden Markov Models, Deep Learning and experiments as well as simulation based methods. The use of HAIT in automotive applications will pave the way to its utilisation in other disciplines such as aviation and maritime.